Though considerable progress has been made in inferring phylogenetic relationships of many plant lineages, deep unresolved nodes remain a common problem that can impact downstream efforts, including taxonomic decision-making and character reconstruction. The Core Goodeniaceae is a group affected by this issue: data from the plastid regions trnL-trnF and matK have been insufficient to generate adequate support at key nodes along the backbone of the phylogeny. We performed genome skimming for 24 taxa representing major clades within Core Goodeniaceae. The plastome coding regions (CDS) and nuclear ribosomal repeats (NRR) were assembled and complemented with additional accessions sequenced for nuclear G3PDH and plastid trnL-trnF and matk. The CDS, NRR, and G3PDH alignments were analyzed independently and topology tests were used to detect the alignments' ability to reject alternative topologies. The CDS, NRR, and G3PDH alignments independently supported a Brunonia (Scaevola s.l. (Coopernookia (Goodenia s.l.))) backbone topology, but within Goodenia s.l., the strongly-supported plastome topology (Goodenia A (Goodenia B (Velleia+Goodenia C))) contrasts with the poorly supported nuclear topology ((Goodenia A+Goodenia B) (Velleia+Goodenia C)). A fully resolved and maximally supported topology for Core Goodeniaceae was recovered from the plastome CDS, and there is excellent support for most of the major clades and relationships among them in all alignments. The composition of these seven major clades renders many of the current taxonomic divisions non-monophyletic, prompting us to suggest that Goodenia may be split into several segregate genera.

Aim The plant family Goodeniaceae includes 12 genera, largely restricted to Australia. They are a diverse and conspicuous element of the country's ecosystems and an important clade to consider in the effort to reconstruct the historical biogeography of this continent. We used a time‐calibrated molecular phylogeny to perform ancestral‐area reconstructions for the family, in order to: (1) determine its area of origin; (2) test whether vicariance played a role in the history of lineages occupying the south‐western and south‐eastern regions; (3) trace diversification processes in the central arid zone; and (4) trace potential eastward dispersals into the Pacific. Location Australia and the Pacific. Methods We employed a previously published molecular dataset to develop a time‐calibrated phylogeny for the Goodeniaceae using beast 1.6.1 with both primary and secondary node calibrations. Ancestral areas were reconstructed via both dispersal–extinction–cladogenesis (DEC) and statistical dispersal–vicariance analysis (S‐DIVA). Results The Goodeniaceae split from Asteraceae + Calyceraceae c. 78 Ma, with the divergence of the Lechenaultia, Anthotium and Dampiera (LAD) clade from the remainder of the family c. 67 Ma. Although the geographical origin of the family is equivocal, the LAD clade originated in south‐western Australia, and the remainder of the family, especially Goodenia s.l., probably originated in the central Eremaean zone. The majority of Goodenia cladogenesis events occurred during the Miocene, with expansions to the north, south‐east and south‐west. The crown radiation of Scaevola s.l. occurred later, during the second half of the Miocene, and has involved several recent extra‐Australian dispersals into the Pacific. Main conclusions Although it is unclear where in Australia the family originated, our results support a south‐western origin for the LAD clade and considerable diversification of some clades within Goodenia s.l. and Scaevola s.l. in that zone. A relative paucity of mid‐Miocene sister relationships between south‐western and south‐eastern lineages suggests that the occupants of these regions were not affected by contemporary vicariance events. Instead, our results suggest that the aridifying Eremaean zone was an important source and sink for lineages, especially in Goodenia s.l. Dispersal is inferred for the extra‐Australian lineages due to their recent origins.

Premise of the study:The use of genome skimming allows systematists to quickly generate large data sets, particularly of sequences in high abundance (e.g., plastomes); however, researchers may be overlooking data in low abundance that could be used for phylogenetic or evo-devo studies. Here, we present a bioinformatics approach that explores the low-abundance portion of genome-skimming next-generation sequencing libraries in the fan-flowered Goodeniaceae.Methods:Twenty-four previously constructed Goodeniaceae genome-skimming Illumina libraries were examined for their utility in mining low-copy nuclear genes involved in floral symmetry, specifically the CYCLOIDEA (CYC)-like genes. De novo assemblies were generated using multiple assemblers, and BLAST searches were performed for CYC1, CYC2, and CYC3 genes.Results:Overall Trinity, SOAPdenovo-Trans, and SOAPdenovo implementing lower k-mer values uncovered the most data, although no assembler consistently outperformed the others. Using SOAPdenovo-Trans across all 24 data sets, we recovered four CYC-like gene groups (CYC1, CYC2, CYC3A, and CYC3B) from a majority of the species. Alignments of the fragments included the entire coding sequence as well as upstream and downstream regions.Discussion:Genome-skimming data sets can provide a significant source of low-copy nuclear gene sequence data that may be used for multiple downstream applications.

Phylogenetic results suggest the American bulb-bearing Oxalis originated in southern South America, dispersed repeatedly to North America, and had multiple transitions from tristyly to distyly. This study adds to our understanding of biogeographic history and breeding system evolution and provides a foundation for more precise inferences about the study group.

Core Goodeniaceae is a clade of ~330 species primarily distributed in Australia. Considerable variation in flower morphology exists within this group and we aim to use geometric morphometrics to characterize this variation across the two major subclades: Scaevola sensu lato (s.l.) and Goodenia s.l., the latter of which was hypothesized to exhibit greater variability in floral symmetry form. We test the hypothesis that floral morphological variation can be adequately characterized by our morphometric approach, and that discrete groups of floral symmetry morphologies exist, which broadly correlate with subjectively determined groups. From 335 images of 44 species in the Core Goodeniaceae, two principal components were computed that describe >98% of variation in all datasets. Increasing values of PC1 ventralize the dorsal petals (increasing the angle between them), whereas increasing values of PC2 primarily ventralize the lateral petals (decreasing the angle between them). Manipulation of these two morphological “axes” alone was sufficient to recreate any of the general floral symmetry patterns in the Core Goodeniaceae. Goodenia s.l. exhibits greater variance than Scaevola s.l. in PC1 and PC2, and has a significantly lower mean value for PC1. Clustering clearly separates fan-flowers (with dorsal petals at least 120° separated) from the others, whereas the distinction between pseudo-radial and bilabiate clusters is less clear and may form a continuum rather than two distinct groups. Transitioning from the average fan-flower to the average non-fan-flower is described almost exclusively by PC1, whereas PC2 partially describes the transition between bilabiate and pseudo-radial morphologies. Our geometric morphometric method accurately models Core Goodeniaceae floral symmetry diversity.